HPLC Analysis
The HPLC analysis is nowadays the routinely applied instrumental technique to analyze flavor compounds in vanilla products or their extracts. HPLC analysis is used to identify and quantify relevant compounds as well as in the further evaluation of authenticity or adulteration. Other techniques such as GC analyses have been published (Mosandl and Scharrer, 2001) but not been established in quality control and authenticity checks. An authenticity check based on the enantio selective analysis of the volatile fraction of vanilla extracts has been investigated (Mosandl and Hener, 2001). Among others, γ-nonalactone is one of the minor chiral compounds in vanilla and was extracted and separated into the R- and S-enantiomer. The low enantiomeric excess of R-γ-nonalactone of 45–63% was judged to be insignificantly high enough to provide an unequivocal proof for adulteration with racemic γ-nonalactone. HPLC methods established for vanilla are robust, rapid, and well suited for routine analysis. Sample preparation of the HPLC analysis is done by extraction in case of the beans, or simply dilution for extracts. The effects of the sample preparation techniques on the analytical results have been investigated (Ehlers et al., 1999). These results indicate that the ratios of typical vanilla ingredients such as vanillin, vanillic acid, p-hydroxybenzaldehyde, and p-hydroxybenzoic acid can vary depending not only on the raw material but also on the sample preparation method used.
Different HPLC methods have been developed and published and finally being adopted as official methods (Taylor, 1993; ISO, 1999) or even being part of the food law in France (Arrêté du juin 11, 1987).
The stationary phases are dominantly reversed-phase materials, mostly RP-18 materials but also RP-8 and others such as alkyl halogen-modified silica gel (Taylor, 1993).
The methods applying reversed-phase separation typically run with methanol/ water mobile phases in isocratic mode or more often with gradient elution at acidic pH values and finally using UV- or DAD-detection (diode array detector). The quantification can be based on internal standards, and also on external calibration as most of the detected compounds are available as pure chemicals. Run times are around 20 min; however, with special columns with selected stationary phases, the run time can be reduced to less than 5 min (Tracy et al., 2008).
Typical parameters for HPLC analysis of vanilla extracts are given in Table 15.1. The compounds that have been identified are also subject to publications and are described elsewhere in this book (see Chapter 12). Typical compounds that are detected with these analyses are vanillin, vanillic acid, p-hydroxybenzaldehyde, and p-hydroxy benzoic acid. These analytes are seen as impact compounds and represent a typical profile (Figure 15.1). A simple authenticity check can be performed based on the presence of untypical compounds such as ethyl vanillin but also on the absence of one or more of these impact compounds. Other minor compounds that are detected by HPLC, such as anisaldehyde, may serve as indicators to distinguish V. planifolia from V. tahitiensis (Ehlers et al., 1994; Oberdieck, 1998).
| Instrument Agilent | 1100 |
| Column | Lichrospher 100 RP18 (5 μm) 124 × 4 mm |
| Injection volume | 2 μL |
| Eluent/gradient | Acetonitrile/water resp. phosphate buffer |
| Detection | Diode array detector |
FIGURE 15.1 HPLC chromatogram of vanilla extract compounds.
Their quantitative distribution is typical for the kind of vanilla product and the extraction technique. Different studies have investigated the profile resulting from these impact compounds in vanilla extracts and found to be specific and therefore suitable to evaluate the authenticity of a vanilla extract (Fayet et al., 1987; Juergens, 1981). To transform the profile of these compounds into a numerical scale, the ratios of their concentration were calculated and limits were suggested. Five ratios of vanilla compounds have been introduced into French law to evaluate the authenticity (République Française, 1988) of vanilla products. In 2003, the ratios have been amended by the Direction Générale de la Concurrence, de la Consommation et de la Répression des Fraudes (DGCCRF) to reflect findings over a longer period of time and additional information from recent crops at that time (Note d’information, 2003). The limits are shown in Table 15.2. This evaluation is suitable only for alcoholic extracts of cured whole vanilla beans. Changes in the solvent composition or the use of alternative extraction agents such as CO2 lead to ratios that do not necessarily comply with the limits given in Table 15.2 (Quirin and Gerard, 1998). Also the source of the vanilla beans used to prepare the extracts influence the profile and herewith the ratios. Further studies have partly disproved the conclusion of fraud when a ratio
| Ratio | Ranges according to DGCCRFa 1988 | Ranges according to DGCCRFb 2003 |
|---|---|---|
| Vanillin/p-hydroxybenzaldehyde | 10–20 | 10–20 |
| Vanillin/p-hydroxybenzoic acid | 53–110 | 40–110 |
| Vanillin/vanillic acid | 15–29 | 12–29 |
| p-hydroxybenzoic acid/p-hydroxybenzaldehyde | 0.15–0.35 | 0.15–0.35 |
| Vanillic acid/p-hydroxybenzaldehyde | 0.53–1.00 | 0.53–1.50 |
a République Française (1988).
b Note d’information (2003).
is out of the ranges given (John and Jamin, 2004; Littmann-Nienstedt and Ehlers, 2005; Gassenmeier et al., 2008) and the suitability of the so-called ratios has been argued. In an information letter, the International Organization of the Flavor Industry (IOFI) has expressed their reservation toward the applicability of the ratios and overestimation of their validity (IOFI Information Letter, 2000) despite the fact that these ratios are sometimes erroneously cited as “IOFI values.” The questionable validity of the ratios has been applied not only to vanilla beans and extracts but also to vanilla flavors and even flavored food. An interpretation of ratios for flavors and food is far beyond of what has seriously been investigated at the scientific level. Thus, the so-called ratios are a precheck for the authenticity of vanilla beans and extracts thereof, in case the conditions of extraction are well known to the evaluator. Other techniques such as stable isotope ratio mass spectrometry (IRMS) or quantitative site-specific nuclear magnetic resonance spectroscopy provide much clearer indication of adulteration of vanilla products (Kempe and Kohnen, 1999).